Osteopontin Enhances the Expression and Activity of
MMP-2 via the SDF-1/CXCR4 Axis in Hepatocellular
Carcinoma Cell Lines
Rihua Zhang1, Xiaolin Pan1, Zuhu Huang2, Georg F. Weber3, Guoxin Zhang1*
1Department of Gastroenterology, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China, 2Department of Infection Diseases, The First Affiliated
Hospital of Nanjing Medical University, Nanjing, China, 3University of Cincinnati Academic Health Center, College of Pharmacy, Cincinnati, Ohio, United States of America
Background and Aims: Osteopontin, SDF-1a, and MMP-2 are important secreted molecules involved in the
pathophysiology of human hepatocellular carcinoma (HCC). This study investigates the effect of the SDF-1a/CXCR4 axis
on expression and activity of MMP-2 induced by osteopontin.
Methods: The expression of CXCR4, SDF-1a, MMP-2 and their associated cellular signaling cascades, involving Akt and MAP
Kinases, were determined by Western blotting. The activities of MMP-2 and MMP-9 were assayed by gel zymography. The
role of the osteopontin receptors integrin avb3 and CD44v6 was evaluated using neutralizing antibodies. We also
established CXCR4-deficient SMMC7721 cell lines by transfection with miRNA-CXCR4 plasmids and determined cell invasion
activity in a transwell assay.
Results: In comparison with untreated cells, recombinant human osteopontin (rhOPN) up-regulated CXCR4, SDF-1a, and
MMP-2 expression about 5-, 4-, and 6-fold on the protein levels through binding to integrin avb3 and CD44v6 in
hepatocellular carcinoma cells (SMMC7721 and HepG2). Inhibition of the SDF-1a/CXCR4 axis down-regulated the rhOPN-
induced MMP-2 expression and activity. rhOPN also activated Akt, p38 and JNK. Down-regulation of CXCR4 decreased the
rhOPN-induced invasion in SMMC7721 cells.
Conclusion: These results indicate that rhOPN up-regulates MMP-2 through the SDF-1a/CXCR4 axis, mediated by binding to
integrin avb3and CD44v6 and activating the PI-3K/Akt and JNK pathways in HepG2 and SMMC7721 cells. Therefore, the
osteopontin-SDF-1a/CXCR4-MMP-2 system may be a new therapeutic target for treating HCC progression.
Citation: Zhang R, Pan X, Huang Z, Weber GF, Zhang G (2011) Osteopontin Enhances the Expression and Activity of MMP-2 via the SDF-1/CXCR4 Axis in
Hepatocellular Carcinoma Cell Lines. PLoS ONE 6(8): e23831. doi:10.1371/journal.pone.0023831
Editor: Jean-Marc Vanacker, Institut de Ge ´nomique Fonctionnelle de Lyon, France
Received March 8, 2011; Accepted July 26, 2011; Published August 31, 2011
Copyright: ? 2011 Zhang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by grants from National Natural Science Foundation of China (No. 81072032 and 30770992), and Social Development Funds
of Jiangsu Province and from Jiangsu Health Department, China (No. B52007070 and H200702). GW is founder and CEO of MetaMol Theranostics. The funders had
no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
Many experimental and clinical studies have demonstrated that
a substantial number of secreted factors are involved in the
pathophysiology of human hepatocellular carcinoma (HCC) .
Among them, the cytokine osteopontin, the SDF-1a/CXCR4 axis
(stromal cell derived factor-1/ CXC chemokine receptor 4), and
MMP enzymes are thought to play key roles in invasion and
Osteopontin is an aspartate-rich protein expressed by various
tissues and cell types. The existence of variant forms of osteopontin,
been described. sOPN interacts with integrins and variant CD44. It
contains several cell binding domains, including an arginine-
glycine-aspartate (RGD)-motif that engages a subset of cell surface
integrins (avb3, avb1, avb5, and a8b1), a serine-valine-valine-
tyrosine-glutamate-leucine-arginine (SVVYGLR)-containing do-
main that interacts with other integrins (a9b1, a4b1and a4b7),
and an ELVTDFTDLPAT domain that has been reported to bind
to integrin a4b1.The CD44-bindingsite hasbeen mappedto the
C-terminal portion of osteopontin. The cytokine activates various
signaling pathways to mediate multiple functions such as inflam-
mation, cell adhesion, migration and tumor invasion. Osteopontin
up-regulates matrix metalloproteinase 2 (MMP-2). In MDA-MB-
231 human breast cancer cells, MMP-2 was significantly decreased
following exposure to an inhibitor of osteopontin . Further study
hasshown that osteopontin activatesthe phosphoinositide 3-kinase/
Akt survival pathway [7,8].
SDF-1 and its receptors, such as CXC chemokine receptor 4
(CXCR4), are thought to play critical roles in motility, homing,
and proliferation of many cancer cells . SDF-1, which belongs
to the CXC chemokine subfamily, is produced in two forms, SDF-
1a (CXCL12a) and SDF-1b (CXCL12b), by alternative splicing of
the SDF-1 gene. The binding of SDF-1a to its receptor CXCR4
stimulates receptor dimerization and activates downstream
PLoS ONE | www.plosone.org1August 2011 | Volume 6 | Issue 8 | e23831
signaling to play an important role in a wide array of disease
We thus assessed the role of the SDF-1a/CXCR4 axis in the
process of OPN mediated MMP-2 up-regulation in the two
cell lines,HepG2 and
Materials and Methods
rhOPN (Recombinant human Osteopontin/his) (#1433-OP/CF)
was purchased from R&D Systems (USA). PD98059 (#9900),
LY294002 (#9901), MAPKFamily Antibody Sampler
(#9926), Phospho-Akt (Ser473), Antibody (#9271) and SDF-1
antibody (#3530) were purchased from Cell Signaling Technology
(USA). Rabbit polyclonal to CXCR4 (#ab2074) was obtained from
(#MAB4073), Anti-integrin aV clone AV1 monoclonal antibody
affinity purified polyclonal antibody (#AB1868P) came from
Millipore (USA). SB203580 (#S8307), SP600125 (#S5567) and
ECM gel (#e1270) were obtained from Sigma-Aldrich (USA).
AMD3100 (#10011332) was purchased from Cayman Chemical
and Functional Grade Purified anti–human CXCR4 (12G5) (#16-
9999) from eBioscience (USA).
The human hepatocellular carcinoma cell lines SMMC7721
and HepG2 cells  were cultured in DMEM supplemented with
10% fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin
sulfate (100 mg/ml), and maintained at 37uC with 5% CO2in a
Construction of miRNA-CXCR4 expression plasmids and
stable clone selection
Four distinct domains within the coding region of the human
CXCR4 cDNA were targeted for RNA interference. For this
purpose, four pairs of reverse complementary oligonucleotides
were designed and synthesized as Table 1.
The oligonucleotides were annealed and inserted into the
#K4936-00) to create pcDNA6.2-GW/EmGFP–miR -CXCR4-
1-4, 2-4, 3-1, and 4-4. A control construct was also created.
We used lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA)
to separately transfect the five kinds of plasmids into SMMC7721
cells. To select for successful transfectants, the cells were cultured
48 hours after transfection in selection medium containing 3 mg/
ml blasticidin (Sigma-Aldrich, Saint Louis, MO, USA). Blasticidin-
resistant cells were maintained in culture medium supplemented
with 3 mg/ml blasticidin for further analysis.
Gel zymography for evaluation of gelatinolytic activity
In this study, the human hepatocellular carcinoma cell lines
SMMC7721 and HepG2 (16106) were seeded in 6-cm (diameter)
dishes containing complete growth medium. After 12 hours
incubation in DMEM with 0.1% BSA, the medium was changed
to DMEM with 0.1% BSA in the absence or presence of rhOPN
(50 nM) for 60 hours. The rhOPN concentration is in the range
commonly associated with cancer . We then collected the
supernatant and centrifuged it at 12,000 rpm for 10 min to pellet
insoluble material. The protein concentration in the supernatant
was determined using a Protein Assay Rapid Kit (Bio-Rad, Osaka,
Japan). Samples containing 40 mg total protein in sample buffer
(10% SDS, 4% sucrose, 0.25 M Tris-HCl, pH 6.8 and 0.1%
bromophenol blue) were used in gelatin zymography. The
samples, diluted 1:1 with 26 sample buffer, were not boiled but
warmed in a water bath (55uC) for 3–5 min before being subjected
to electrophoresis in a 10% SDS-polyacrylamide gel (SDS-PAGE)
containing 0.1% gelatin under non-reducing conditions. The gel
was washed twice for 30 min in 2.5% Triton X-100 at room
temperature to remove the SDS. After the second wash, all but 2–
3 ml of the Triton X-100 was removed, and 100 ml of
development buffer (0.05 M Tris-HCl pH 8.8, 5 mM CaCl2,
0.02% NaN3, 0.02% Brij) was added for further incubation for
24 hours at 37uC. The gel was then stained for 3 hours in
Coomassie blue (0.1% Coomassie brilliant blue R250 (w/v) in
fixing/destaining solution) and destained in fixing/destining
solution (methanol: acetic acid: water, 4.5:1:4.5) until clear bands
of gelatinolysis appeared on a dark background. Total activity was
analyzed using a scanning densitometer with molecular analysis
software (Bio-Rad) .
Enzyme-linked immunosorbent assay (ELISA) was done with a
human SDF-1 Quantikine kit (R&D), used in accordance with the
manufacturer’s protocol. In this study, the human hepatocellular
carcinoma cell lines SMMC7721 and HepG2 (16106) were seeded
in 6-cm (diameter) dishes containing complete growth medium.
After 12 hours incubation in DMEM with 0.1% BSA, the medium
was changed to DMEM with 0.1% BSA in the absence or
Table 1. Reverse complementary oligonucleotides.
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org2August 2011 | Volume 6 | Issue 8 | e23831
presence of rhOPN (50 nM) for 24, 48, 72 hours. We then
collected the supernatants and measured total protein content
using the BCA protein assay kit (Pierce) before analysis. Results are
representative of three independent experiments.
Western blotting analysis
The SMMC7721 and HepG2 cells (16106) were treated with
rhOPN (50 nM) for 48 hours, and then lysed in RIPA buffer.
Equal amounts of protein (60 mg) were electrophoresed on 12%
SDS-PAGE gels and electrophoretically transferred to Immobilon-
P membranes (Millipore, Bedford, MA, USA). The membranes
were probed overnight at 4uC with antibody to CXCR4 (1:1000),
MMP-2, SDF-1a and monoclonal anti-a-tubulin (1:5000) in
TBST containing 1% BSA (w/v). The blots were then incubated
for 2 hours with anti-rabbit or anti-mouse secondary antibodies,
the immune complex was detected using an ECL plus detection kit
(Pierce, Rockford, IL, USA), and analyzed using a scanning
densitometer with molecular analysis software (Bio-Rad).
Integrin avb3and CD44v6 neutralization
SMMC7721 and HepG2 cells were cultured as described above
in the presence of anti-integrin avb3or anti-CD44v6 neutralizing
antibodies, or of control IgG. After 60 hours, the cells were
collected and Western blotting was performed for the relevant
Cell invasion assay
Cell invasion was studied using 24-well transwell plates (Corning
Costar, Schiphol-Rijk, Netherland). 60 ml of the ECM gel solution
was added to the top compartment of each cell culture insert and
dried overnight under laminar air flow. The cells under study were
harvested, washed twice with PBS, resuspended in serum-free culture
medium with 0.2% BSA and adjusted to a final concentration of 106
per ml. 600 ml serum-free DMEM/ 0.2% BSA containing rhOPN
(50 nM) was added to the lower compartment of each well, and
200 ml of the cell suspension was added to the pre-coated upper
compartment. The plate with inserts was incubated for 48 hours in a
cell culture incubator at 37uC and 5% CO2. To determine the
background migration, some wells of the 24-well plate were prepared
without rhOPN in the lower compartment. Cells remaining on the
top side of the filter were removed by soft mechanical dislodging, and
the number of cells migrating to the bottom of the filter was counted
using a light microscope (in each chamber, six fields were counted at
2006magnification for each condition).
SMMC7721 and HepG2 cells were collected with trypsin/
EDTA, washed with fluorescence-activated cell sorting (FACS)
buffer (phosphate-buffered saline [PBS], 2 mM EDTA, and 0.5%
BSA), and then incubated in FACS buffer for 1 hour at 4uC in the
presence of monoclonal antibodies at the manufacturer’s recom-
Figure 1. SDF-1a, CXCR4 and MMP-2 expression are induced by rhOPN in SMMC7721 and HepG2 cells. SMMC7721 cells (A) or HepG2
cells (B) were stimulated with various concentrations of rhOPN for 48 hours, the cells were collected, and SDF-1a, CXCR4 and MMP-2 were detected
by Western blotting assay. (C) HepG2 cells were stimulated with 50 nM rhOPN for increasing time frames, the cells were collected, and SDF-1a, CXCR4
and MMP-2 were detected by Western blotting assay. (D) SDF-1 ELISA of culture supernatants (SMMC7721 and HepG2) after 0–72 hours of rhOPN
(50 nM). (E) MMP-2 activity was analyzed by gelatin zymography after stimulation with 50 nM rhOPN for 60 hours in the SMMC7721 and HepG2 cell
lines. *denotes P,0.05 versus control. The results presented are representative of at least three independent experiments.
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org3 August 2011 | Volume 6 | Issue 8 | e23831
mended concentrations. Binding of anti-CD44var (v6) and anti-
integrin aV clone AV1 were visualized with FITC-conjugated
rabbit anti-mouse immunoglobulin (Chemicon, Temecula, CA).
The cells were washed, fixed with 1% paraformaldehyde and the
fluorescence was quantified on 10,000 cells using a FacsCalibur
with Cellquest software (BD Biosciences, PharMingen).
The data were analyzed by two-tailed Student’s t-test for single
comparisons and by one-way analysis of variance for multiple
group comparisons. Differences were considered significant at a
probability of error below 5% versus control.
Osteopontin up-regulates SDF-1a, CXCR4 and MMP-2
expression in hepatocellular carcinoma cells
To determine the effect of rhOPN on the SDF-1a/CXCR4
axis and MMP-2 expression, Western blotting analysis and gel
zymography were done in two human hepatocellular carcinoma
cell lines, SMMC7721 and HepG2. Figure 1A shows that the
expression of SDF-1a, CXCR4, and MMP-2 protein were
induced by rhOPN. There was an apparent increase in the
CXCR4 protein level when the concentration of rhOPN was
3.12 nM in SMMC7721 cells, and the same phenomenon was
also observed in HepG2 cells (Figure 1B). Figure 1C shows that
MMP-2 expression was detectable within 24 hours after the
addition of rhOPN, reached a maximum around 60 hours. The
MMP-2 levels increased in a time-dependent manner in HepG2
cells. SDF-1a and CXCR4 expression increased accordingly
(Figure 1C and 1D). Figure S1A and S1B are quantification of
expression described in Figure 1A and 1B based on grayscale
After identifying MMP-2 protein expression in SMMC7721 and
HepG2 cells, we further analyzed the MMP-2 activity in the two
cell lines by gelatin zymography. The results demonstrated that
the activity of MMP-2 but not MMP-9 was induced by rhOPN at
a dose of 50 nM (Figure 1E).
Figure 2. Effects of the SDF-1a/CXCR4 axis on rhOPN-induced MMP-2 expression and activity. (A) Verification by Western blotting of the
miRNA knockdown of CXCR4 showed a significant reduction of the CXCR4 protein in all clones (1-4, 2-4, 3-1, 4-4). After blocking the SDF-1a/CXCR4
axis with miRNA-CXCR4 and inhibitors (SDF-1 neutralizing antibody at 100 ng/ml, CXCR4 inhibitor 12G5 at 50 mg/ml, or CXCR4 inhibitor AMD3100 at
500 ng/ml), the cells were stimulated by rhOPN in serum-free medium for 60 hours, the cells were collected and analyzed by Western blotting in
SMMC7721 cells (C) and in HepG2 cells (F). The supernatants of SMMC7721 cells (B) and HepG2 cells (E) were analyzed by gelatin zymography. (D)
and (G) show the densitometric ratio of MMP-2 protein/a-tubulin. (H) Western blotting was used to assay the MMP-2 expression induced by rhOPN
(50 nM) or/and SDF-1 (30 nM) for 48 hours. * denotes P,0.05 versus control. The results presented are representatives of at least three independent
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org4 August 2011 | Volume 6 | Issue 8 | e23831
The SDF-1a/CXCR4 axis is involved in osteopontin-
induced MMP-2 expression and activity
The MMPs are a large family of proteolytic enzymes, which play
an important role in cancer invasion and metastasis due to their
Among them, MMP-9 and MMP-2 have been found to be highly
associated with metastatic spread by various cancers. Therefore, to
determine whether the SDF-1a/CXCR4 axis mediates osteopon-
tin-induced MMP-2 expression and activity, we established
CXCR4-deficient SMMC7721 cell lines (clone 1-4, 2-4, 3-1 and
4-4) through the transfection of miRNA-CXCR4. SMMC7721-
vector was used as a control. The CXCR4 protein was detected by
Western blotting. CXCR4 expression was significantly down-
regulated in all of the four miRNA clones (Figure 2A).
The SMMC7721 cells and the miRNA transfectant clones were
stimulated with rhOPN for 60 hours. At that time, the cells and
their conditioned medium were collected for gelatin zymography
and Western blotting. Decreased amounts of MMP-2 proteins were
detected in CXCR4-deficient SMMC7721 cells (clones 1-4, 2-4, 3-1
and 4-4) (Figure 2C), compared to SMMC7721 and vector control.
The increased activity of MMP-2 but not MMP-9 was abolished in
the CXCR4-deficient SMMC7721 cells (Figure 2B).
Figure 3. Integrin avb3and CD44 mediated OPN-induced CXCR4 expression in SMMC7721 and HepG2 cells. FACS analysis using
monoclonal antibodies to avb3integrin (left) and CD44 (right) was done for SMMC7721 cells (A) and HepG2 cells (C), stimulated by rhOPN for
24 hours. The grey area represents isotype control, while the dark line represents the control and the grey line represents the experimental group.
SMMC7721 (E) and HepG2 (F) cells were treated with rhOPN (50 nM), in the presence of neutralizing antibodies to integrin avb3or CD44v6, or control
IgG. After 60 hours, the cells were collected and Western blotting was performed to detect CXCR4. (B), (D), (G) and (H) are quantitative evaluations.
The results are shown as mean 6 standard deviation (n=3). * denotes P,0.05 compared to rhOPN treatment in the absence of antibody.
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org5 August 2011 | Volume 6 | Issue 8 | e23831
To further elucidate the role of the SDF-1a/CXCR4 axis in
human hepatocellular carcinoma, we detected the expression and
activity of MMP-2 induced by 50 nM rhOPN in the presence or
absence of SDF-1a neutralizing antibody, CXCR4 inhibitor
12G5, or CXCR4 inhibitor AMD3100. Firstly, we have assessed
the toxicity of AMD3100, and the results showed that AMD3100
(500 ng/ml) had no effect on proliferation of HepG2 (figure S2A).
The results indicated that exposure of the HepG2 cells to anti-
SDF-1 antibody, CXCR4 inhibitor 12G5, or CXCR4 inhibitor
AMD3100 decreased the rhOPN-induced MMP-2 activity
(Figure 2E) and expression (Figure 2F and D). In order to confirm
that SDF-1 actually mediates the observed OPN effects on MMP-
2 expression, Western blotting was used to assay the MMP-2
expression levels induced by rhOPN (50 nM) or/and SDF-1
(30 nM) for 48 hours. The results indicate that SDF-1 does
generate the same response as OPN in terms of MMP-2
expression, but rhOPN and SDF-1 have no synergistic effects on
MMP-2 expression (Figure 2H), presumably because they belong
to the same pathway.
Osteopontin up-regulates CXCR4 in hepatocellular
carcinoma cells through both major receptors
Known receptors for osteopontin through which the cytokine is
thought to influence diverse physiological and pathological processes
respective roles of the cognate receptors in osteopontin-dependent
signaling pathways, we initially performed flow cytometry analysis in
SMMC7721 and HepG2 cells for the expression levels of avb3integrin
and CD44. CD44 was expressed a high levels in both cell lines,
regardless of whether they had been cultured in the presence or
absence of rhOPN. Integrin avb3was about twofold inducible by
rhOPN from very low baseline expression levels in SMMC7721 cells
(Figure 3A and B). The receptor was only marginally inducible by
rhOPN in HepG2 cells (Figure 3C and D).
Neutralizing antibodies to integrin avb3and CD44v6 were used to
further test whether these osteopontin receptors are involved in the
observed induction of CXCR4 expression. Both antibodies down-
regulated the CXCR4 expression induced by rhOPN, about 4.2- and
1.8-fold in SMMC7721 cell (Figure 3E and G), 3.2- and 2.2- fold in
HepG2 cell (Figure 3F and H). It is known that CD44 and integrin b3
can interact. The contribution by both receptors suggests that rhOPN
may engage a CD44v6/integrin avb3complex in the cancer cell
membrane,activating downstreamsignal transduction pathways in the
hepatocellular carcinoma cells HepG2 and SMMC7721.
Osteopontin-induced CXCR4 and MMP-2 expression are
mediated by PI-3K/Akt and JNK
Osteopontin has been reported to activate various kinases such
as PI-3K, protein kinase C, and the MAP Kinases, which have
three major subgroups (ERK, p38, and JNK). We first asked
whether osteopontin activates Akt (the downstream target of PI-
3K) and MAPK in SMMC7721 and HepG2 cells. As shown in
Figure 4A–D, Akt, p38, and JNK phosphorylation were
stimulated within 30 min following the addition of rhOPN,
whereas ERK1/2 was not. Figure S3A, B, C and D are
quantification of expression described in Figure 4A, B, C, and D
based on grayscale analysis (analyzed from three independent
experiments, *P,0.05 versus control. The data are representative
of three experiments).
To define the role of the PI-3K/Akt and MAPK pathways in
rhOPN-induced SDF-1, CXCR4 and MMP-2 expression, we used
inhibitors for PI-3K and MAPKs. SDF-1, CXCR4 and MMP-2
Figure 4. The rhOPN-induced expression of CXCR4 and MMP-2 depends on the PI3K/Akt and JNK pathways. rhOPN induced the
phosphorylation of Akt (A), p38 (B) and JNK (C), but not ERK1/2 (D) in SMMC7721 and HepG2 cells. The cells (16106cells/ml) were left untreated or
stimulated with rhOPN (50 nM) for 30 min and total cell lysates were subjected to Western blotting analysis. After pretreatment of SMMC7721 cells
(E) or HepG2 cells (F) with PD98059 (ERK inhibitor, 100 mM), SB203680 (p38 inhibitor, 100 mM), SP600125 (JNK inhibitor, 100 mM), LY294002 (PI-3K
inhibitor, 100 mM) or DMSO for 45 min, the cells were treated with rhOPN (50 nM) for 48 hours and total cell lysates were subjected to Western
blotting analysis for MMP-2 or CXCR4.
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org6 August 2011 | Volume 6 | Issue 8 | e23831
expression, stimulated by rhOPN, were decreased significantly in
the presence of LY294002 (PI-3K inhibitor) or SP600125 (JNK
inhibitor). However, as shown in Figure 4E and F (figure S3E and
F are quantification of expression described in Figure 4E and F
based on grayscale analysis), the inhibitors of ERK or p38 had
little effect on the rhOPN-induced CXCR4 expression in
SMMC7721 and HepG2 cells. These results indicate that
osteopontin-mediated CXCR4 and MMP-2 expression depends
on activation of the PI-3K/Akt and JNK pathways.
CXCR4 is required for osteopontin-induced cell invasion
The observation that rhOPN induced SDF-1a, CXCR4 and
MMP-2 expression in human hepatocellular carcinoma cells
suggested that the SDF-1a/CXCR4 axis might play a role in
osteopontin-dependent tumor progression. Therefore, we assessed
the invasive response to rhOPN in a transwell assay. As shown in
Figure 5, rhOPN stimulated the invasion response of SMMC7721
cells (6-fold) and SMMC7721-miRNA-CXCR4 cells, clones 1-4
and 2-4 (2-fold). Importantly, invasiveness was decreased in clone
1-4 and 2-4 cells, about four fold lower than in SMMC7721cells.
These results indicate that rhOPN activates cell invasion through
the induction of CXCR4 in SMMC7721 cells.
Hepatocellular carcinoma (HCC) is one of the most common
and malignant neoplasms worldwide. Its pathophysiology is
associated with multiple cytokines and secreted factors, including
osteopontin, SDF-1 and its receptor CXCR4, as well as MMP-2
and MMP-9 [16,17,18]. Osteopontin expression is up-regulated in
tumors and blood of human HCC patients compared to healthy
controls [19,20,21]. It has been suggested that osteopontin
overproduced by tumor cells may act as a potent angiogenic
factor . Our study indicates that osteopontin stimulates MMP-
2 expression and activity through a hitherto undefined pathway.
Both MMP-2 and MMP-9 play important roles in the
pathogenesis of many cancers [19,20]. Our results are consistent
with previous results showing that osteopontin up-regulates MMPs
[21,22,23]. It has been reported that the osteopontin-induced
activation of MMP-2 or MMP-9 is mediated by the PI-3K/Akt/
NF-kB signaling pathway [21,24,25]. Osteopontin may promote
the activation of pro-MMP-9, but not MMP-2, through an
NADPH oxidase-associated signaling cascade . While we
found no rhOPN effect on MMP-9, our study has identified a
novel pathway to MMP-2 expression and activation, which is
mediated by the SDF-1a/CXCR4 axis (Figure 6). Our results
(Figures 4) further demonstrate that the p38 and ERK pathways
are involved in the expression of MMP-2, but not SDF-1 and
CXCR4 expression, induced by rOPN. In this study, our focus is
on the OPN-dependent enhancement of the expression and
activity of MMP-2 via the SDF-1/CXCR4 axis. p38 and ERK
MAPK induces MMP-2 expression in many cells, for example,
baicalein downregulates the protein expression levels of MMP-2
by inhibiting the expression of p-Akt, p-ERK, p-p38 and p-JNK
; the down-regulation of p38 MAPK and JNK by siRNA
transfection resulted in a decrease in MMP-2 expression by
MelJuso cells . The rhOPN-induced CXCR4 expression is
dependent on CD44 and integrin receptors, and is regulated by
the PI-3K/Akt and JNK pathways (Figures 3 and 4) in the two
hepatocellular carcinoma cell lines (HepG2 and SMMC7721)
tested. Moreover, decreased production of CXCR4 and MMP-2
in association with lower invasion of hepatocellular carcinoma cells
could be related to the down-regulation of metastasis [28,29]. Our
data add to this evidence (Figure 5), indicating that osteopontin,
CXCR4, and MMP-2 are key cytokines for HCC progression.
Although our results are derived from two model cell lines,
published evidence corroborates that they are relevant for human
Osteopontin and CXCR4 have been used as markers for
immune activation  for homing precursor cells [32,33] and for
metastasizing cancer cells [34,35,36]. Further, osteopontin and
CXCR4 may serve as early biomarkers for cancer detection .
However, reports in the literature have not yet provided a
functional link between these molecules. The identification of
CXCR4 as a downstream target of osteopontin and an essential
mediator in the induction of MMP-2 closes this gap. Osteopontin
up-regulates MMP-2 through activating the SDF-1a/CXCR4
Figure 5. The rhOPN-induced SMMC7721 cells invasion is
mediated by CXCR4. The invasion assay was set up in transwell
chambers. Cell culture inserts with 8.0-mm pore diameter were used to
separate the top and bottom chambers. 60 ml of ECM gel solution was
added to the upper compartment of each cell culture insert and dried
overnight under laminar air flow. SMMC7721 parent cells, vector
controls, and CXCR4 miRNA clones (1-4, 2-4) were plated onto the
membrane of the top chamber, and rhOPN was administered to the
bottom chamber. After 48 hours, the cells that had invaded to the
lower surface of the membrane were enumerated. (A) Bright-field
image of cells migrated to the bottom of chambers on the inserts (2006
original magnification). (B) Quantification of cell invasion. The open
bars represent no osteopontin, the filled bars represent rhOPN
treatment. In each chamber, six fields were counted at 2006
magnification for each condition by two investigators. * indicates
P,0.05 versus control. The data are representative of three experi-
Figure 6. Model for the mechanism of osteopontin-dependent
MMP-2 up-regulation in hepatocellular carcinoma. The results of
the present study show that osteopontin up-regulates SDF-1a, CXCR4,
and MMP-2 via integrin avb3and CD44v6, as well as PI-3K/Akt and JNK.
These data are consistent with an osteopontin-induced autocrine loop
of SDF-1a/CXCR4 activation that leads to tumor invasion, in part via
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org7August 2011 | Volume 6 | Issue 8 | e23831
axis, mediated by binding to integrin avb3 and CD44v6 and
activating the PI3K/Akt and JNK pathways in hepatocellular
carcinoma cells (HepG2 and SMMC7721). Therefore, the
osteopontin-SDF-1a/CXCR4- MMP-2 system may be a promis-
ing therapeutic target.
expression described in Figure 1A and 1B based on
grayscale analysis (analyzed from three independent
experiments). *denotes P,0.05 versus control.
Figure S1A and S1B are quantification of
as 1, and cell numbers collected at all other time points
The cell numbers at 12 h postplating were set
were compared with the initial values at 12-h time point.
Results were expressed as the mean6SD. * P,0.05 when
compared with the DMSO control.
Figure 4A, B, C, D, E and F based on grayscale analysis
(analyzed from three independent experiments). *P,0.05
versus control. The data are representative of three experiments.
Quantification of expression described in
Conceived and designed the experiments: GZ GW. Performed the
experiments: RZ XP. Analyzed the data: GW GZ. Contributed
reagents/materials/analysis tools: ZH. Wrote the paper: RZ GZ.
1. Ogawa K (2009) Molecular pathology of early stage chemically induced
hepatocarcinogenesis. Pathol Int 59: 605–622.
2. Ramaiah SK, Rittling S (2008) Pathophysiological role of osteopontin in hepatic
inflammation, toxicity, and cancer. Toxicol Sci 103: 4–13.
3. Gazitt Y (2004) Homing and mobilization of hematopoietic stem cells and
hematopoietic cancer cells are mirror image processes, utilizing similar signaling
pathways and occurring concurrently: circulating cancer cells constitute an ideal
target for concurrent treatment with chemotherapy and ntilineage-specific
antibodies. Leukemia 18: 1–10.
4. Lunxiu Q, Zhaoyou T (2002) The prognostic molecular markers in
hepatocellular carcinoma. World J Gastroenterol 8: 385–392.
5. Scatena M, Liaw L, Giachelli CM (2007) Osteopontin: a multifunctional
molecule regulating chronic inflammation and vascular disease. Arterioscler
Thromb Vasc Biol 27: 2302–2309.
6. Mi Z, Guo H, Russell MB, Liu Y, Sullenger BA, et al. (2009) RNA aptamer
blockade of osteopontin inhibits growth and metastasis of MDA-MB231 breast
cancer cells. Mol Ther 17: 153–161.
7. E1-Tanani M (2008) Role of osteopontin in cellular signaling and metastatic
phenotype. Front Biosci 13: 4276–4284.
8. Junko M, Shigeyuki K, Yutaka M, Toshimitsu U (2010) Osteopontin; as a target
molecule for the treatment of inflammatory diseases. Curr Drug Targets 22:
9. Jaszczynska-Nowinka K, Markowska A (2009) New cytokine: stromal derived
factor-1. Eur J Gynaecol Oncol 30: 124–127.
10. Miroslaw J (2009) Functional diversity of SDF-1 splicing variants. Cell Adh Migr
11. Tiveron M-C, Cremer H (2008) CXCL12/CXCR4 signalling in neuronal cell
migration. Curr Opin Neurobiol 18: 237–244.
12. Kang Y, Chen BJ, Deoliveira D, Mito J, Chao NJ (2010) Selective enhancement
of donor hematopoietic cell engraftment by the CXCR4 antagonist AMD3100
in a mouse transplantation model. PLoS One 5: 113–116.
13. Veldkamp CT, Seibert C, Peterson FC, De la Cruz NB, Haugner JC, et al.
(2008) Structural basis of CXCR4 sulfotyrosine recognition by the chemokine
SDF-1/CXCL12. Sci Signal 1: ra4.
14. Brakora KA, Lee H, Yusuf R, Sullivan L, Harris A, et al. (2004) Utility of
osteopontin as a biomarker in recurrent epithelial ovarian cancer. Gynecol
Oncol 93: 361–365.
15. Clark IM, Hawkes SP, Li H, Taniguchi GT (1990) Methods in Molecular
Biology Humana Press.
16. Xiang Z-L, Zeng Z-C, Tang Z-Y, Fan J, Zhuang P-Y, et al. (2009) Chemokine
receptor CXCR4 expression in hepatocellular carcinoma patients increases the
risk of bone metastases and poor survival. BMC Cancer 9: 176.
17. Wang B, Hsu SH, Majumder S, Kutay H, Huang W, et al. (2010) TGFbeta-
mediated upregulation of hepatic miR-181b promotes hepatocarcinogenesis by
targeting TIMP3. oncogene 29: 1787–1797.
18. Takafuji V, Forgues M, Unsworth E, Goldsmith P, Wang XW (2007) An
osteopontin fragment is essential for tumor cell invasion in hepatocellular
carcinoma. oncogene 26: 6361–6371.
19. Levicar N, Nuttall RK, Lah TT (2003) Proteases in brain tumour progression.
Acta Neurochir (Wien) 145: 825–838.
20. Philip S, Bulbule A, Kundu GC (2004) Matrix metalloproteinase-2: mechanism
and regulation of NF-kappaB-mediated activation and its role in cell motility and
ECM-invasion. Glycoconj J 21: 429–441.
21. Lai CF, Seshadri V, Huang K, Shao JS, Cai J, et al. (2006) An osteopontin-
NADPH oxidase signaling cascade promotes pro-matrix metalloproteinase 9
activation in aortic mesenchymal cells. Circ Res 98: 1479–1489.
22. Ye S (2006) Influence of matrix metalloproteinase genotype on cardiovascular
disease susceptibility and outcome. Cardiovasc Res 69: 636–645.
23. Hu T, Luan R, Zhang H, Lau WB, Wang Q, et al. (2009) Hydrogen peroxide
enhances osteopontin expression and matrix metalloproteinase activity in aortic
vascular smooth muscle cells. Clin Exp Pharmacol Physiol 36: 626–630.
24. Song G, Ouyang G, Mao Y, Ming Y, Bao S, et al. (2009) Osteopontin promotes
gastric cancer metastasis by augmenting cell survival and invasion through Akt-
mediated HIF-1alpha up-regulation and MMP9 activation. J Cell Mol Med 13:
25. Yan W, Qian C, Zhao P, Zhang J, Shi L, et al. (2010) Expression pattern of
osteopontin splice variants and its functions on cell apoptosis and invasion in
glioma cells. Neuro Oncol 12: 765–775.
26. Ling Wang YL, Yan Chen (2010) Flavonoid baicalein suppresses adhesion,
migration and invasion of MDA-MB-231 human breast cancer cells. Cancer
Letters 297: 42–48.
27. Hong I-K, Kim Y.-M, Lee H (2005) Tetraspanin CD9 induces MMP-2
expression by activating p38 MAPK, JNK and c-Jun pathways in human
melanoma cells. Experimental and Molecular Medicine 37: 230–239.
28. McKenna GJ, Chen Y, Smith RM, Meneghetti A, Ong C, et al. (2002) A role
for matrix metalloproteinases and tumor host interaction in hepatocellular
carcinomas. Am J Surg 183: 588–594.
29. Liu H, Pan Z, Li A, Fu S, Lei Y, et al. (2008) Roles of chemokine receptor 4
(CXCR4) and chemokine ligand 12 (CXCL12) in metastasis of hepatocellular
carcinoma cells. Cell Mol Immunol 5: 373–378.
30. Budhu AS, Zipser B, Forgues M, Ye QH, Sun Z, et al. (2005) The molecular
signature of metastases of human hepatocellular carcinoma. Oncology 69:
31. Jakobsen LP, Borup R, Vestergaard J, Larsen LA, Lage K, et al. (2009)
Expression analyses of human cleft palate tissue suggest a role for osteopontin
and immune related factors in palatal development. Exp Mol Med 41: 77–85.
32. Otsuru S, Tamai K, Yamazaki T, Yoshikawa H, Kaneda Y (2008) Circulating
bone marrow-derived osteoblast progenitor cells are recruited to the bone-
forming site by the CXCR4/stromal cell-derived factor-1 pathway. Stem Cells
33. Kollet O, Dar A, Shivtiel S, Kalinkovich A, Lapid K, et al. (2006) Osteoclasts
degrade endosteal components and promote mobilization of hematopoietic
progenitor cells. Nat Med 12: 657–664.
34. Javelaud D, Mohammad K, McKenna CR, Fournier P, Luciani F, et al. (2007)
Stable overexpression of Smad7 in human melanoma cells impairs bone
metastasis. Cancer Res 67: 2317–2324.
35. Lin HM, Chatterjee A, Lin YH, Anjomshoaa A, Fukuzawa R, et al. (2007)
Genome wide expression profiling identifies genes associated with colorectal liver
metastasis. Oncol Rep 17: 1541–1549.
36. Gaviraghi M, Tunici P, Valensin S, Rossi M, Giordano C, et al. (2010)
Pancreatic cancer spheres are more than just aggregates of stem marker positive
cells. Biosci Rep Epub ahead of print.
37. Rogers CD, Fukushima N, Sato N, Shi C, Prasad N, et al. (2006) Differentiating
pancreatic lesions by microarray and QPCR analysis of pancreatic juice RNAs.
Cancer Biol Ther 5: 1383–1389.
OPN Induces MMP-2 through the SDF-1/CXCR4 Axis
PLoS ONE | www.plosone.org8 August 2011 | Volume 6 | Issue 8 | e23831